Silhouette-based gait recognition using Procrustes shape analysis and elliptic Fourier descriptors

Abstract

This paper presents a gait recognition method which combines spatio-temporal motion characteristics, statistical and physical parameters (referred to as STM-SPP) of a human subject for its classification by analysing shape of the subject's silhouette contours using Procrustes shape analysis (PSA) and elliptic Fourier descriptors (EFDs). STM-SPP uses spatio-temporal gait characteristics and physical parameters of human body to resolve similar dissimilarity scores between probe and gallery sequences obtained by PSA. A part-based shape analysis using EFDs is also introduced to achieve robustness against carrying conditions. The classification results by PSA and EFDs are combined, resolving tie in ranking using contour matching based on Hu moments. Experimental results show STM-SPP outperforms several silhouette-based gait recognition methods.

Clothing and carrying condition invariant gait recognition based on rotation forest

“…Model-based methods (e.g., [17,23,9]) use a structural model to measure time-varying gait parameters, e.g., gait period, stance width and stride length, and a motion model to analyse the kinematical and dynamical motion parameters of the subject, e.g., rotation patterns of hip and thigh, and joint angle trajectories, to obtain gait signatures. The modelfree gait recognition methods in [4,3,27] analyse the dynamic and/or static gait characteristics of silhouettes or the extreme outer boundary of silhouettes, i.e., contours of a gait sequence. The performance of these methods largely depends on the correctness of the background segmentation techniques, presence of occluding objects in the scene and shadows under feet, as these factors considerably determine the quality of the silhouettes and the extracted contours.…”

“…RCK-G is robust to limited clothing variation, but sensitive to carrying conditions. The methods in [3,4,20] aim to achieve invariance to carrying conditions. The method based on spatio-temporal motion characteristics, statistical and physical parameters (STM-SPP) [3] analyses the shape of a contour using Procrustes analysis at the double support phase and elliptic Fourier descriptors (EFDs) at ten phases of a gait period.…”

“…The methods in [3,4,20] aim to achieve invariance to carrying conditions. The method based on spatio-temporal motion characteristics, statistical and physical parameters (STM-SPP) [3] analyses the shape of a contour using Procrustes analysis at the double support phase and elliptic Fourier descriptors (EFDs) at ten phases of a gait period. The method in [4] combines modelbased and model-free approaches to analyse the spatio-temporal shape and dynamic motion (STS-DM) characteristics of a subject's contour.…”

“…The components of swing phase are: pre-swing, mid-swing and terminal swing. A detailed description of these phases are provided in [3].…”

“…Several state-of-the-art gait recognition methods [4,27,3] analyse the dynamic and/or static gait characteristics of silhouettes or the extreme outer boundary of silhouettes, i.e., contours of a gait sequence for identifying a human subject. The performance of these methods largely depends on the correctness of the background segmentation techniques, presence of occluding objects in the scene and shadows under feet, as these factors considerably determine the quality of the silhouettes and the extracted contours.…”